Georgia Tech to Build $20M National AI Supercomputer
The National Science Foundation (NSF) has awarded Georgia Tech and its partners $20 million to develop a new supercomputer named Nexus. This system is designed to accelerate scientific research through advanced artificial intelligence (AI) capabilities.
Nexus will be among the most advanced AI-focused research tools in the United States. It aims to support scientists working on urgent challenges such as new drug development, clean energy solutions, brain research, and manufacturing innovations.
Overview
Georgia Tech is no stranger to high-performance computing; it already hosts the PACE Hive Gateway supercomputer. Nexus will expand this capacity, combining AI with powerful computing resources to enable faster and deeper scientific discovery.
Ángel Cabrera, president of Georgia Tech, emphasized the institution's role in providing AI talent and technology, noting that hosting Nexus supports a new wave of AI-centered innovation nationwide.
Features of Nexus
Built specifically for AI research, Nexus will offer a user-friendly interface accessible to researchers across the country. It will support diverse fields including climate science, healthcare, aerospace, and robotics.
Katie Antypas, director of the NSF’s Office of Advanced Cyberinfrastructure, highlighted Nexus’s novel approach, which combines persistent scientific services with traditional high-performance computing, accelerating scientific workflows.
Nexus Supercomputer — In Simple Terms
- Built for the future of science: Nexus is engineered to handle demanding AI research tasks, from disease cures to quantum materials engineering.
- Blazing fast: It performs over 400 quadrillion operations per second, comparable to every person on Earth performing 50 million calculations every second nonstop.
- Massive brain plus memory: Equipped with 330 trillion bytes of memory, Nexus can process complex problems and vast datasets.
- Storage: Featuring 10 quadrillion bytes of flash storage — equivalent to 10 billion reams of paper stacked to reach 500,000 km, enough to stretch from Earth to the moon and back one-third of the way.
- Supercharged connections: Lightning-fast data transfer minimizes wait times for researchers.
- Open to U.S. researchers: Scientists from any U.S. institution can apply for access through an NSF review process.
Why Now?
AI is changing how scientific research is conducted, enabling analysis of massive datasets and modeling complex systems faster than before. Yet, many institutions lack access to the computing power required for such work.
Nexus addresses this gap by making top-tier AI infrastructure available nationwide, not just at tech hubs. This levels the playing field for researchers across the country.
Suresh Marru, principal investigator of Nexus and director of Georgia Tech’s Center for AI in Science and Engineering (ARTISAN), notes that Nexus is designed to make powerful AI tools easier to use and more widely accessible.
Srinivas Aluru, Regents’ Professor and senior associate dean at Georgia Tech’s College of Computing, added that Nexus marks Georgia Tech’s entry into the league of academic supercomputing centers. The project builds on earlier NSF-funded initiatives like the HIVE supercomputer and leverages Georgia Tech’s CODA data center.
A National Collaboration
Georgia Tech is collaborating with the National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign, which operates several leading academic supercomputers. Together, they will link Nexus with NCSA’s systems over a high-speed network, creating a national research infrastructure.
Charles Isbell, chancellor of the University of Illinois and former dean at Georgia Tech’s College of Computing, described Nexus as a symbol of what institutions can achieve when working together to advance science and society.
What’s Next
Construction of Nexus will begin this year, with completion expected by spring 2026. Afterward, researchers can apply for access through an NSF review process. Georgia Tech will manage the system, provide user support, and reserve up to 10% of Nexus’s capacity for its own research needs.
Vivek Sarkar, dean of computing at Georgia Tech, emphasized that Nexus will enable researchers to tackle some of today’s most challenging problems and inspire discoveries that are yet to be imagined.
For researchers interested in expanding their AI skills to make the most of resources like Nexus, exploring Complete AI Training’s latest AI courses can be a valuable step.
Your membership also unlocks:
